Invariably, when the subject of the Crash of 1987 is discussed, something called portfolio insurance is mentioned in association with the decline. To this day few investors are familiar with what it is, and most only have a vague notion that it had something to do with exacerbating the steep fall that occurred in the index averages almost a decade ago. Portfolio insurance was a risk management weapon forged in the raging furnace of market decline, but when tested in battle it quickly revealed a critical flaw.
The concept of portfolio insurance, indeed the name itself, reflects the ardent wishes of its creators for a utopian investment vehicle capable of minimizing the pain of investment loss. The idea was simple, all one would have to do was pay something akin to an insurance premium, and investment downside could be forever limited. Unfortunately, as so often occurs in the realm of human affairs, reality served an unwilling partner.
The inspiration for portfolio insurance oddly enough came from an academic setting. In September 1976, Hayne Leland, a young finance professor at Berkeley, consumed with worry over his familys finances had a midnight brainstorm. He approached a colleague, Mark Rubinstein, the following morning and the pair became so enamored of their ideas that they decided to form a company to market their new product called portfolio insurance. However, it wasnt until a full two years later that the duo had ironed out all the kinks in their trading strategy, and with the help of professional marketer and portfolio theorist John OBrien they scored their first client in 1980.
From inauspicious beginnings came a virtual tidal wave of demand for the groups product, as well as a whole host of competitors offering similar investment insurance. In fact, by 1987 approximately $60 billion in equity assets were covered by different varieties of portfolio insurance, the majority of which was on behalf of conservative pension funds.
The initial concept underlying Lelands creation was to replicate the performance of a put option in what was referred to as a dynamically programmed system, where a client was automatically shifted out of a position (via computer) when it began to fall, increasing the clients cash as long as the stock fell and insuring that a predetermined amount was all that the client could lose.
Heres how it worked: Say you bought 100 shares of Replicants R Us at $100 a share, and simultaneously bought a put -- the option but not the obligation to sell Replicants stock to someone else at a stated price over a specified period of time -- with an exercise price of $90. In this way, no matter how low Replicants R Us shares dropped due to malfunctioning replicants killing humans, you couldnt lose more than $10 on each share. In this example the floor was set at $90, and by the time the stock reached this price in the market, your portfolio would be 100% in cash. The distance then, between the starting point and the floor, was like a deductible on a normal insurance policy -- the $10 per share being the amount that you as a policyholder would have to cover. So if the market fell, the portfolio would slowly liquidate some positions, but still hold some stock. If the market rose, the portfolio would be buying, but still hold some cash. The slight degree of underperformance, in both directions, that results serves as the premium for the insurance.
Today, in a period when program trading is routine, this particular scheme looks a little archaic. However, at the time, the mechanics of running an operation in which hundreds of simultaneous buy and sell orders needed to be executed for just one portfolio were understandably complicated as well as costly. In addition, active portfolio managers bristled at being second-guessed by "some computer. The saving grace for human and computer alike came in the form of futures contracts on the S&P 500, made available in 1983. These contracts allowed investors to buy or sell a proxy for the overall market, that is, an index of 500 leading companies in leading industries.
Under the guise of portfolio insurance, the futures transaction in this instance was an agreement to sell the intangible commodity of 500 stocks in the index at a specified date and a prearranged price. Portfolio managers would hedge their stock portfolios by selling index futures. Instead of buying and selling hundreds of stocks and options, the S&P index futures promised an effective, inexpensive, and greatly simplified form of portfolio insurance. They functioned almost exactly like the put scenario described previously, with one crucial difference that left the promise unfulfilled.
The owner of the futures contract makes a cash settlement based upon the variation in the index between the signing of the contract and its maturity. Every day investors are required to fork over cash to the exchange (which represents the other side of the transaction, as opposed to an individual or firm) to keep the contracts fully collateralized. The actual construction of the futures contract was not the problem in 1987, it was the selling of shares in the stock portfolio in order to make the hedge worthwhile.
Recall the put scenario. This option was "put" to the seller at the exercise price of $90. Again, that means however low Replicants R Us shares dropped in the market, the seller of the Replicants put is legally bound to buy the shares at $90 if the owner of the option puts the stock. That was the insurance -- Relicants R Us stock could fall to $30, but the seller was obligated to buy at $90! However, with the stock index future, it could be sold and the money could be collected, but in order to make it all worthwhile the stocks in the portfolio had to be sold to prevent the loss.
The crucial difference boiled down to the fact that the market was not obligated to purchase any of that portfolio stock at the price that the portfolio manager wanted. Those who used index futures as a hedge assumed that the necessary liquidity would be available from the market. It wasn't. The truth is, during the crash most "insured" portfolios did better than those that were "uninsured," but the selling of the portfolio stock took place at prices far lower than anticipated. For example, instead of locking in a price at $90 for Replicants R Us shares, those that insured their portfolios with index futures would have undoubtedly received a price below $90 because the buyers were just not there during the waves of selling that occurred.
On the weekend before the crash, an enormous overhang of sell orders had built up, for in the previous week the Dow Jones Industrial Average had fallen 250 points, or about 10% (with half of the drop on Friday). During all this, managers of "insured" portfolios were selling index futures and dumping their portfolio stock. On Monday all the programmed sales kicked in, the market dropped 100 point by noon, another 200 points in the ensuing two hours, and 300 points in the final hour. As Peter Bernstein notes in Against the Gods: The Remarkable Story of Risk, "The cost of portfolio insurance in that feverish market turned out to be much higher than paper calculations predicted."
Back to: 1987 Revisited